Abstract Dementia is caused by a constellation of age-associated neurodegenerative diseases collectively termed Alzheimer’s disease and related dementias (ADRD). ADRDs impair individuals' ablity to correctly function and constitutes a major worldwide public health problem, with no effective treatments available. ADRDs affect people of all ethnicities, but Hispanics are especially affected by it. Hispanics have 1.5-fold higher risk of developing ADRD than non-Hispanic whites. ADRDs are complex phenotypes and their diagnosis often requires invasive measures. Therefore, identifying reliable noninvasive early ADRD biomarkers is a key topic to provide better early diagnosis and guide patient’s treatment. Toward this goal, this project aims to identify novel non-invasive endophenotypes using transcriptomic profiling in an existing cohort of Hispanics from the Maracaibo Aging Study (MAS) living in Maracaibo, Venezuela. The MAS, begun in 1998, has surveyed longitudinally genetic, environmental, and socioeconomic elements that may contribute to ADRD prevalence. The MAS project currently includes ~5000 elderly individuals (aged ≥55 yr) and a subset of 415 genetically related participants that have been extensively phenotyped, including complete MRI brain imaging and comprehensive cognitive assessment with a high ADRD prevalence (6.8%). Using a family-based design, we calculate that genetic factors account for 67% of the variability in disease risk. To identify novel ADRD and cognitive decline endophenotypes, we will construct a RNASeq library for 1000 MAS participants, and apply our ERV (Endophenotype Ranking Value) method. We will divided MAS participants into two equal subsets of 500 each for discovery and replication sets respectively coming from Santa Rosa and Santa Lucia communities in Venezuela. The discovery subset will include 122 ADRD cases defined by DSM-IVR criteria and, more importantly, a set of 243 1st- to 5th-degree biological relatives who are not currently suffering from dementia. We will contrast them with a set of 135 unaffected individuals (lacking known relatives with ADRD) to identify gene expression endophenotypes. Differences between these groups can only occur due to genetic correlation between ADRD risk and an endophenotype as identified by the ERV method taking into account covariates such as age, age at onset, sex and APOE locus status. Endophenotypes detected in the discovery set will be confirmed as biomarkers in a set of 250 ADRD cases and 250 unaffected controls using a linear mixed model and later annotated by means of a scientific literature review. We tested our approach's validity using existing data from our Genetics of Brain Structure Study where we identified strong associations between known the expression of candidate genes’ and dementia risk. We will apply the ERV method to the MAS gene expression data and predict that it will perform even better since the MAS population manifests higher ADRD prevalence. This proj...